Throughout this application, various publications are referenced by Arabic numerals. Full citations for these publications may be found at the end of the specification immediately preceding the claims. The disclosure of these publications is hereby incorporated by reference into this application to describe more fully the state of the art to which this invention pertains.
The life cycle of animal viruses is characterized by a series of events that are required for the productive infection of the host cell. The initial step in the replicative cycle is the attachment of the virus to the cell surface which is mediated by the specific interaction of the viral attachment protein (VAP) to receptors on the surface of the target cell. The pattern of expression of these receptors is largely responsible for the host range and tropic properties of viruses. The interaction of the VAP with cellular receptors therefore plays a critical role in infection and pathogenesis of viral diseases and represents an important area to target the development of anti-viral therapeutics.
Cellular receptors may comprise of all of the components of membranes, including proteins, carbohydrates, and lipids. Identification of the molecules mediating the attachment of viruses to the target cell surface has been made in a few instances. The most extensively characterized viral receptor protein is CD4 (T4) (1). CD4 is a nonpolymorphic cell surface glycoprotein that is expressed primarily on the surface of helper T lymphocytes, cells of the monocyte/macrophage lineage and dendritic cells. CD4 associates with major histocompatibility complex (MHC) class II molecules on the surface of antigen-presenting cells to mediate efficient cellular immune response interactions. In humans, CD4 is also the target of interaction with the human immunodeficiency virus (HIV).
HIV primarily infects helper T lymphocytes, monocytes, macrophages and dendritic cells--cells that express surface CD4. HIV-infected helper T lymphocytes die, and the loss of these CD4+ T lymphocytes is one marker of the progress of HIV infection. The depletion of these cells is probably an important cause of the loss of immune function which results in the development of the human acquired immune deficiency syndrome (AIDS). In contrast to helper T lymphocytes, other CD4+ cells, notably dendritic cells, monocyte and macrophages, are chronically infected by HIV. They produce virus over a long period of time and appear to be major reservoirs of virus in vivo (2, 3).
The initial phase of the HIV replicative cycle involves the high affinity interaction between the HIV exterior envelope glycoprotein gp120 and surface CD4 (Kd approximately 4.times.10.sup.-9 M) (4). Several lines of evidence demonstrate the requirement of this interaction for viral infectivity. In vitro, the introduction of a functional cDNA encoding CD4 into human cells which do not express CD4 is sufficient to render otherwise resistant cells susceptible to HIV infection (5). In vivo, viral infection appears to be restricted to cells expressing CD4. Following the binding of HIV gp120 to cell surface CD4, viral and target cell membranes fuse, resulting in the introduction of the viral nucleocapsid into the target cell cytoplasm.
Characterization of the interaction between HIV gp120 and CD4 has been facilitated by the isolation of cDNA clones encoding both molecules (6, 7). CD4 is a nonpolymorphic, lineage-restricted cell surface glycoprotein that is a member of the immunoglobulin gene superfamily. High-level expression of both full-length CD4 and truncated, soluble versions of CD4 (sCD4) have been described in stable expression systems. The availability of large quantities of purified sCD4 has permitted a detailed understanding of the structure of this complex glycoprotein. Mature CD4 has a relative molecular mass (Mr) of 55 kilodaltons and consists of an amino-terminal 372 amino acid extracellular domain containing four tandem immunoglobulin-like regions denoted V1-V-4, followed by a 23 amino acid transmembrane domain and a 38 amino acid cytoplasmic segment. The amino-terminal immunoglobulin-like V1 domain bears 32% homology with kappa light chain variable domains. Three of the four immunoglobulin-like domains contain a disulphide bond (V1, V2 and V4), and both N-linked glycosylation sites in the carboxy-terminal portion of the molecule are utilized (4, 8).
Experiments using truncated sCD4 proteins demonstrate that the determinants of high-affinity binding to HIV gp120 lie within the V1 domain (9-11). Mutational analysis of V1 has defined a discrete gp120 binding site (residues 38-52) of the mature CD4 protein) that comprises a region structurally homologous to the second complementarity-determining region (CDR2) of immunoglobulins (11). The production of large quantities of V1V2 has permitted a structural analysis of the two amino-terminal immunoglobulin-like domains. The structure determined at 2.3 angstrom resolution reveals that the molecule has two tightly associated domains containing the immunoglobulin-fold connected by a continuous beta strand. The putative binding sites for monoclonal antibodies, class II MHC molecules and HIV gp120 (as determined by mutational analysis) map on the molecular surface (12, 13).
A number of therapeutic strategies have been proposed using CD4-based molecules to target HIV or HIV-infected cells which express gp120. These strategies have the advantage that they depend on the interaction between CD4 and gp120. This interaction is essential for virus infection, so CD4-based strategies should be effective against all strains of HIV. Moreover, it is highly unlikely that escape mutants would develop with mutations in gp120 which eliminate CD4 binding. This is in marked contrast with therapeutic strategies which target other regions of gp120 (e.g. vaccine approaches) or other viral proteins (e.g. reverse transcriptase) where the therapy is effective against a limited subset of HIV strains, and/or the virus can mutate and become resistant to the therapy.
In one example of CD4-based therapies, a soluble version of the entire extracellular segment of CD4 (V1-V4, termed sCD4) has been developed (14). In vitro experiments demonstrate that: 1) SCD4 acts as a "molecular decoy" by binding to HIV gp120 and inhibiting viral attachment to and subsequent infection of human cells; 2) sCD4 "strips" the viral envelope glycoprotein gp120 from the viral surface (although this is more important with laboratory isolates of HIV than with clinical isolates of the virus); and 3) sCD4 blocks the intercellular spread of virus from HIV-infected cells to uninfected cells by inhibiting virus-mediated cell fusion (1, 15).
In addition to in vitro results, experiments with sCD4 in simian immunodeficiency virus (SIV)-infected rhesus monkeys have been described. These studies demonstrated that administration of sCD4 to SIV-infected rhesus monkeys leads to a diminution of the viral reservoir.
Phase I human clinical trials with sCD4 demonstrated that there is no significant toxicity or immunogenicity associated with administration of sCD4 at doses as high as 30 mg/day. Preliminary antiviral studies were inconclusive with respect to CD4 cell count and levels of HIV antigen (16, 17).
Although these in vitro, primate and human studies with sCD4 produced encouraging results, they also defined some limitations. In particular, the measured serum half-life of sCD4 is very short (30-45 minutes in humans following intravenous administration (16, 17)). It is hard to imagine that sCD4 administration could eliminate HIV from the body, but rather it would be used to delay or prevent the spread of infection and the development of disease. Therefore a therapeutic regimen might involve regular treatment with the protein. However, the short half-life of sCD4 might make it difficult to maintain sufficient levels in the plasma to give a therapeutic effect. This problem is compounded by the fact that much higher levels of sCD4 are required to neutralize clinical isolates of HIV compared to laboratory isolates, although all clinical isolates can be neutralized at some concentration (18). To make a CD4-based molecule with a longer half-life, several groups have now made chimeric CD4-based molecules which comprise the gp120 binding region of CD4 and another protein such as an immunoglobulin molecule. These molecules are described in greater detail below.
Another drawback to sCD4 is that it does not kill HIV-infected cells such as monocytes/macrophages and dendritic cells. These cells act as reservoirs for HIV and chronically produce virus which infects other cells such as helper T lymphocytes. The CD4-based chimeras mentioned above may also have limited efficacy in killing HIV-infected cells. While a chimera between CD4 and human immunoglobulin gamma 1 may kill HIV infected cells by antibody-dependent cellular cytotoxicity (ADCC) in vitro, experience with anti-tumor monoclonal antibodies suggests that monoclonal antibody-mediated ADCC is rarely effective in vivo. Therefore, another CD4-based approach has been developed where sCD4 is linked to a toxin molecule. These chimeras can bind to, and kill, HIV-infected cells which express gp120 on their surfaces.
In one study, sCD4 was coupled to the deglycosylated A chain of ricin which inactivates ribosomes, therefore inhibiting protein synthesis and killing the cell. This fusion protein was reported to lyse cells infected with five different isolates of HIV, but was nontoxic to uninfected cells (19).
In another study, the V1V2 domains of CD4 were coupled to domains II and III of Pseudomonas exotoxin A (sCD4-PE40) (20). This toxin also blocks protein synthesis, in this case by inactivating elongation factor 2. The sCD4-PE40 fusion protein bound to, and inhibited protein synthesis in, cells expressing the HIV envelope glycoprotein gp120 (20). It has been shown that the sCD4-PE40 conjugate kills cells infected with both laboratory and clinical HIV isolates. This is in contrast to the fact that sCD4 and other CD4-based molecules are much less effective at neutralizing clinical isolates than laboratory isolates of HIV (18). The mechanism for the difference is susceptibility of primary and laboratory isolates appears to be that sCD4 strips gp120 from the virions of laboratory isolates much more efficiently than from clinical isolates (21). However, the resistance to stripping of gp120 in clinical isolates is an asset when using sCD4-toxin molecules to target HIV-infected cells in vivo.
Further studies of CD4-PE40 have shown that this conjugate is capable of eliminating HIV from cultures of infected cells when used in combination with the reverse transcriptase blocker AZT (22). This effect has now been seen with laboratory and clinical isolates, as well as in a variety of different cell types.
In yet another study, a fragment of Diphtheria toxin was genetically fused with the V1 and V2 domains of CD4 (23). This toxin also acts by inactivating elongation factor 2. The CD4-diphtheria toxin fusion protein was effective and specific in killing HIV-infected cells. HIV-infected cultures became resistant to the CD4-Diphtheria conjugate after long term treatment (18 days) for unknown reasons. The significance of this observation is unclear, as the phenomenon was not seen with the other toxin conjugates. Moreover, the CD4-diphtheria toxin study has only been performed using a laboratory isolate of HIV and it will be important to assess its activity against primary clinical isolates, as well as in other cell types.
These CD4-toxin conjugates have some major drawbacks. First, being based on sCD4 or a smaller fragment of CD4, the half-life of the molecules are very short, resulting in a need for higher and more frequent doses than would otherwise be the case. A second drawback is that the toxin moieties are foreign proteins which are highly immunogenic. The development of a strong immune response to the conjugate limits the number of repeat treatments which might be used in one patient. In a similar context, it has been suggested that immunosuppressive agents will have to be administered together with monoclonal anti-tumor antibody-toxin conjugates for tumor therapy (24). However, in the case of HIV infections where the immune system is already compromised, this approach may not be viable.
New families of CD4-based molecules which are toxic to HIV-infected cells are provided in the subject invention. These molecules have many advantages for use in HIV-infected patients to destroy cells which chronically produce HIV, thereby slowing or halting the progress of HIV infections and AIDS. Moreover, the molecules might also be of value in blocking the initial infection of some individuals, for example in babies born of HIV-positive mothers, or in the case of health workers exposed to HIV-positive body fluids. It is likely that transmission in these cases is mainly cell--cell in mechanism, and that killing the infected cells shortly after they enter the target individual could limit or prevent infection.
These CD4-based molecules are based on the conjugation of a non-peptidic toxin or a cytotoxic radioactive moiety with fusion proteins consisting of portions of CD4 and portions of a human immunoglobulin molecule of the gamma 2 subclass. These molecules have considerable advantages over all previously described CD4-based molecules.
The properties of immunoglobulins make them a suitable "backbone" for these CD4-based cytotoxic molecules. Immunoglobulins, or antibodies, are the antigen-binding molecules produced by B lymphocytes which comprise the humoral immune response. The basic unit of an immunoglobulin molecule consists of two identical heavy chains and two identical light chains. The amino-terminus of each chain contains a region of variable amino acid sequence (variable region). The variable regions of the heavy and light chains interact to form two antigen binding sites. The carboxy-terminus of each chain contains a region of constant amino acid sequence (constant domain). The light chain contains a single constant domain, whereas the heavy chain constant domain is subdivided into four separate domains (CH.sub.1, hinge, CH.sub.2, and CH.sub.3). The heavy chains of immunoglobulin molecules are of several types, including mu (M), delta (D), gamma (G), alpha (A) and epsilon (E). The light chains of immunoglobulin molecules are of two types, either kappa or lambda. Within the individual types of heavy and light chains exist subtypes which may differ in effector function. An assembled immunoglobulin molecule derives its name from the type of heavy chain that it possesses.
The development of monoclonal antibodies has circumvented the inherent heterogeneity of antibodies obtained from serum of animals or humans. However, most monoclonal antibodies are derived from cells of mouse origin and therefore are immunogenic when administered to humans. More recent developments combining the techniques of molecular genetics with monoclonal antibody technology has lead to the production of "humanized" chimeric antibodies in vitro. In these chimeric antibodies, the variable domains of human immunoglobulin heavy and light chains are replaced with specific heavy and light chain variable domains from a murine monclonal antibody (25-27). The result of this genetic manipulation is a molecule with specificity for a particular antigen and the characteristics of human immunoglobulins.
Sequence and structural analyses of CD4 indicate that the four extracellular domains are immunoglobulin-like. Since the Fc portion of immunoglobulins controls the rate of catabolism of the molecules (serum half-life ranging from 14 to 21 days) and provides various effector functions, several reports describe the replacement of variable and constant domains of immunoglobulins with the immunoglobulin-like domains of CD4 (21-24).
CD4-IgGl heavy chain fusion proteins resulting in chimeric gamma1 heavy chain dimers have been described (28). These molecules contain the gamma1 heavy chain CH1 domain in addition to the hinge, CH2 and CH3 domains. However, heavy chain assembly and secretion from mammalian cells is less efficient if the CH1 domain is expressed in the absence of light chains (32). Subsequently, a CD4-IgG1 heavy chain fusion protein lacking the CH1 domain and the first five amino acids of the hinge region was described which was secreted to high levels (29).
CD4-IgG1 fusion proteins have also been described. Here the V1-V2 domains of CD4 were fused to the CH1, hinge, CH2 and CH3 domains of a gamma1 heavy chain, and the V1V2 domains of CD4 were fused to the constant region of a kappa light chain (33). CD4-IgM heavy chain fusion proteins have also been described (34).
These fusion proteins have been successfully used to block HIV infection in vitro, and in one case to block the infection of Chimpanzees by a laboratory strain of HIV. As expected, the CD4-immunoglobulin chimeras have a much longer half-life in vivo than does sCD4. As discussed above, however, it is unlikely that these molecules can destroy HIV-infected cells in patients who are already infected with HIV. Their efficacy against primary isolates of HIV has yet also to be established.
These fusion proteins retain various effector functions of immunoglobulin molecules, such as Fc preceptor binding, cell-mediated transfer via an Fc receptor-dependent mechanism and complement activation (29). While these effector functions might have utility in some therapeutic regimens, they are a disadvantage in the present context of developing cytotoxic drugs consisting of toxins or radionuclides linked to CD4-immunoglobulin chimeras.
Many of the functions of antibodies are mediated through their interaction with Fc receptors. These receptors are found on a variety of cells including macrophages, other leucocytes, platelets and placental trophoblasts (35). The Fc receptor binds to the Fc portion of immunoglobulins and the complex can trigger a variety of responses depending on cell type. In the case of macrophages, the response can include phagocytosis and ADCC. With placental trophoblasts, IgG1 binding leads to transfer of the antibody to the fetus.
Human cells express a number of different Fc receptors which are specific for different immunoglobulin isotypes. Three types of human Fc receptor have been described which bind human IgG (Fc.gamma.RI, Fc.gamma.RII and Fc.gamma.RIII) (35). Fc.gamma.RI has a much higher affinity for monomeric IgG than do Fc.gamma.RII and Fc.gamma.RIII. The rank order of activity of Fc.gamma.RI for IgG isotypes is IgG1=IgG3=IgG4. IgG2 does not bind to this receptor. Fc.gamma.RII binds IgG1 and IgG3 more strongly than IgG2 or IgG4. Fc.gamma.RIII recognized only IgG1 and IgC3 (35).
A cytotoxic molecule with FcR-binding capability may kill FcR-bearing cells in an indiscriminate manner. To construct a CD4-based molecule which specifically kills HIV-infected cells, it would be ideal to base it on IgG2 which exhibits little or no FcR binding. Moreover, human IgC2 antibodies exhibit minimal allotypic variation while human IgC1 antibodies have considerable variation. Therefore, to avoid potential immunogenic responses to recombinant molecules containing immunoglobulin domains, a molecule which is the least polymorphic was chosen.
The CD4-IgG2 molecules have advantages relative to the CD4-IgC1 heavy chain dimers which have been described previously. They are also superior to the CD4-toxin molecules which have been developed in the past. Specifically, a CD4-gamma2 chimeric heavy chain homodimer was constructed which contains the V1V2 domains of CD4 and which is efficiently assemble intracellularly and efficiently secreted from mammalian cells as a homodimer, enabling high recovery and purification from the medium of cells expressing this chimeric heavy chain homodimer. To construct this homodimer, the entire hinge, CH2, and CH3 domains from a human gamma2 heavy chain were used, resulting in a chimeric molecule containing the constant domains of a human IgG2 molecule responsible for dimerization and efficient secretion. This is in contrast to the heavy chain dimers described by Capon and Gregory (36) which include the CH1 domain in the CD4-IgG1 heavy chain dimer, resulting in poor secretion and recovery from cell culture medium of the recombinant molecule. Also included is the entire hinge domain of gamma2 heavy chain in the CD4-gamma2 chimeric heavy chain homodimer of this invention to provide efficient dimerization, since the cysteine residues contained in this domain are responsible for forming the disulphide links to the second chain of the homodimer, positioning the two chains in the correct spatial alignment and facilitating formation of the antigen combining site.
In addition to the CD4-gamma2 chimeric heavy chain homodimers, CD4-IgC2 heavy chains were also constructed, which contain the V1V2 domains of CD4 fused to the CH1, hinge, CH2 and CH3 domains of human gamma2 heavy chain. CD4-kappa chimeric light chains were also constructed which contain the V1 and V2 domains of CD4 fused to the entire constant domain of human kappa light chains. When these vectors are co-expressed, they produce a heterotetramer comprising two CD4-IgG2 chimeric heavy chains and two CD4-kappa chimeric light chains. Producing heavy chains which contain the CH1 domain enables efficient association with the CD4-kappa chimeric light chains, resulting in efficient secretion of a CD4-IgC2 chimeric heterotetramer. These CD4-IgG2 chimeric heterotetramers possess increased serum half-lives and increased avidity for HIV as compared with heavy chain dimers.
These CD4-gamma 2 chimeric heavy chain dimers and CD4-IgG2 chimeric heterotetramers are linked to non-immunogenic toxic moieties. Two classes of cytotoxic conjugates have been invented. In the first class, the dimers or tetramers are linked to a non-protein toxin.
One example of this toxin is a member or derivative of the enediyne family of anti-tumor antibiotics, including calicheamicin, esperamicins or dynemicins (37, 38). These toxins are extremely potent and act by cleaving nuclear DNA, leading to cell death. Unlike protein toxins which can be cleaved in vivo to give many inactive but immunogenic polypeptide fragments, toxins such as calicheamicin, esperamicins and other enediynes are small molecules which are essentially non-immunogenic. These non-peptide toxins are chemically-linked to the dimers or tetramers by techniques which have been previously used to label monoclonal antibodies and other molecules. These linking technologies include site-specific linkage via the N-linked sugar residues present only on the Fc portion of the conjugates. Such site-directed linking methods have the advantage of reducing the possible effects of linkage on the binding properties of the CD4 portion of the conjugate.
The second class of cytotoxic conjugates consists of the dimers or tetramers linked to a radionuclide which produces cytotoxic radiation. Examples of the radionuclides which are used include .beta.-particle and .alpha.-particle emitters such as .sup.125 I, .sup.131 I, .sup.90 Y and .sup.212 Bi.
These isotopes are chemically-linked to the dimers or tetramers by techniques which have been used successfully to label monoclonal antibodies and other molecules. These linking technologies include random labeling and site-directed labeling. In the latter case, the labeling is directed at specific sites on the dimer or tetramer, such as the N-linked sugar residues present only on the Fc portion of the conjugates.
In previous studies, anti-tumor antibodies labelled with these isotopes have been used successfully to destroy cells in solid tumors as well as lymphomas/leukemias in animal models, and in some cases in humans (39). The radionuclides act by producing ionizing radiation which causes multiple strand breaks in nuclear DNA, leading to cell death. The isotopes used to produce therapeutic conjugates typically produce high energy .alpha.- or .beta.-particles which have a short path length. Such radionuclides kill cells to which they are in close proximity, for example HIV-infected cells to which the conjugate has attached or has entered. They have little or no effect on neighboring cells. Radionuclides are essentially non-immunogenic.
Both classes of cytotoxic dimer and tetramer conjugates described above have several advantages over other therapeutics which have been described for use against HIV infections. They have a CD4-based mode of action which permits the targeting of all HIV strains and prevents the selection of viral escape mutants. Like other CD4-based molecules, the conjugates may exhibit snyergism when used in combination with other anti-HIV drugs such as AZT. Being conjugated to fragments of IgG2, the molecules have much longer half-lives in vivo than do sCD4-based molecules. They also have the advantage of being dimeric or tetrameric, which increased the avidity of binding to HIV-infected cells. The conjugates kill HIV-infected cells, thereby reducing the rate of spread of HIV infection in vivo, or eliminating infection entirely. All components of the conjugates have been selected for minimal immunogenicity. Being based on IgG2, the conjugates bind minimally, if at all, to Fc receptors, thereby reducing non-specific cell killing.
One use of these radio-conjugates is in the therapy of HIV infections as discussed supra. However, another important application is the use of similar conjugates to detect and localize HIV-infected cells in patients. In this case the conjugates are linked to a .gamma.-radiation emitting isotope such as .sup.111 In, .sup.131 I or .sup.99m Tc. These isotopes emit .gamma.-radiation which passes through tissues for detection/imaging purposes, but causes little ionization or cell death. In the case of an isotope such as .sup.131 I, both high energy .beta.-particles and .gamma.-radiation are produced. This isotope can be used in therapeutic or imaging contexts, depending on the number of .sup.131 I atoms attached to each dimer or tetramer (the specific activity), which governs the cytotoxicity of the dimer or tetramer. Lower specific activities are used for imaging purposes. Such isotopes have been used to image mouse erythroid tumors using leukemia cell-specific monoclonal antibodies labeled with bifunctional radioactive metal chelates (48).
Radioconjugates for diagnostic/imaging purposes would be of value in clinical research to understand the course of HIV infections, as well as in clinical diagnostic applications. For example, imaging could be done in conjunction with treatment using the toxin-conjugated or cytotoxic radionuclide-conjugated dimers and tetramers.
The CD4gamma2 chimeric heavy chain homodimer or CD4-IgG2 chimeric heterotetramer have advantages as imaging agents when compared with antibodies to HIV, since CD4 binds the envelope glycoprotein of all HIV strains with high affinity, whether the envelop glycoprotein is present on the surface of HIV or an HIV-infected cell.